Semianalysis agrees with my analysis of SpaceX getting up to $300 Billion per year by end of 2027.
Spot prices are higher. The AI lease rates make sense.
Coreweave and Nebius are spreading out fixed premium GPUs annual prices but requiring 5 year lockin. 5 years for $100 billion but not available during the next 6-12 months when most valuable.
SpaceX is willing to get premium rents now and willing to replace clients with other clients on the spot market or to make revenue themselves (Via cursor and Grok).
SpaceX And XAI build faster. 1 year to build and dropping to 6-9 months and they are already the lowest cost and getting more cost advantage.
Brian Wang is a Futurist Thought Leader and a popular Science blogger with 1 million readers per month. His blog Nextbigfuture.com is ranked #1 Science News Blog. It covers many disruptive technology and trends including Space, Robotics, Artificial Intelligence, Medicine, Anti-aging Biotechnology, and Nanotechnology.
Known for identifying cutting edge technologies, he is currently a Co-Founder of a startup and fundraiser for high potential early-stage companies. He is the Head of Research for Allocations for deep technology investments and an Angel Investor at Space Angels.
A frequent speaker at corporations, he has been a TEDx speaker, a Singularity University speaker and guest at numerous interviews for radio and podcasts. He is open to public speaking and advising engagements.
Even SpaceX’s permissive data center policies are facing energy capacity reality.
Says Co-Pilot:
“Texas Halts New Data Center Approvals Amid Grid and Water Concerns
Texas Governor Greg Abbott has ordered a pause on all new data center approvals until state regulators complete a comprehensive audit of each project seeking to connect to the state’s power grid Yahoo+1. The move comes as the state faces record electricity demand and growing local opposition to large-scale data center developments.
Why the Pause Was Ordered
Abbott directed the Public Utility Commission of Texas (PUCT) and the Electric Reliability Council of Texas (ERCOT) to verify that every data center project meets strict requirements before grid connection is approved Yahoo+1. These include:
Detailed power and water usage plans
Information on tax incentives and ownership
Measures to mitigate local environmental and community impacts TechCrunch
The audit is intended to protect grid reliability and resilience, especially as ERCOT’s interconnection queue now includes 474 gigawatts of proposed demand—about 90% from data centers—more than five times Texas’ record peak load Reuters+1.
Grid and Energy Context
ERCOT has already seen record-breaking demand in 2026, with a July peak of 91.1 gigawatts, up 6% from the previous record Yahoo. While Texas has expanded renewable capacity and kept electricity prices relatively low, the influx of data centers and crypto-mining facilities is pushing prices higher TechCrunch. Industry forecasts still see Texas surpassing Virginia as the largest U.S. data center market by 2030 Reuters.”
Maybe SpaceX can keep increasing gas turbines – damn the pollution – fast enough to offset grid demand but it’ll cost them both in on-site generation and in newer, more energy and water efficient chips to replace what’s already installed. The typical lifespan of an Nvidea chip is 3 years and SpaceX’s data centers are about 1 year into that already, meaning they’ll have to at least have the next batch on order in 18 months, Q4 2027. And they’ll want to do that even earlier because Nvidea promises 100% re-circulation of water with their next generation chips, which will save 40% on power too because the data centers won’t need to power water chillers either, it’s all air-chilled. What’s to stop other companies finally getting their own data centers up and running, maybe in colder states with less power demands for cooling?
Given that all the other revenue sources are minor and not growing as fast, it’s hard to justify even August 4th’s market cap of $1.6t. I was surprised how little SpaceX’s rocket business contributes to its bottom line, even with Starlink, given how that’s what is promoted as its main mission.
Oops, should have said “Texas’s permissive data policies,” not SpaceX’s.
lifespan of chips is not 3 years! typically 5-6 and now we’re seeing old chip rental rates going up when the come off of contract so now they last even longer.
Water usage being high is hype. Compared to growing some food it is nothing.
Spacex doesn’t do much with grid connections because connecting is so slow. Texas’ order was about grid connection so it doesn’t affect them.
End of 2027 is less than 1.5 years from now. I think 300B per year in 2027 is not gonna happen. People would like to see bright future and exploiting that with some claims is not nice.
Generally overpromising, hyping and manipulating people with false hope is not a fair practice.
Brian just nailed their q2 and underprojected on the December run rate. I think Brian had ~$90b bull case and Johnson (CFO) says $100b arr by December!
Semianalysis’s Patel doesn’t own stocks he covers so no conflict of interest.